Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply the Efficient Market Hypothesis (Fama, 1970) to evaluate information incorporation in asset prices across weak, semi-strong, and strong forms. Use this skill when the user needs to assess market efficiency, determine if a trading strategy can generate abnormal returns, evaluate event studies, or when they ask 'can technical analysis work', 'does the market already know this', or 'is this anomaly exploitable'.
.claude/skills/asgard-ai-platform-grad-emh/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 17% | 0% |
The Efficient Market Hypothesis (Fama, 1970) posits that asset prices fully reflect available information, making it impossible to consistently earn abnormal returns. EMH is organized into three forms — weak, semi-strong, and strong — each defined by the information set reflected in prices.
IRON LAW: In an efficient market, prices reflect available information —
beating the market consistently requires either superior information
or accepting more risk. No free lunch.Key assumptions:
| Form | Information Reflected | Implication | |------|----------------------|-------------| | Weak | Historical prices | Technical analysis cannot earn excess returns | | Semi-strong | All public info | Fundamental analysis cannot earn excess returns | | Strong | All info (public + private) | Even insiders cannot earn excess returns |
Any test of efficiency is simultaneously a test of the asset pricing model used to define "abnormal" return.
markdown## EMH Assessment: [Market / Strategy] ### Efficiency Form Tested - Form: [weak / semi-strong / strong] - Information set: [description] ### Evidence | Test | Result | Supports Efficiency? | |------|--------|---------------------| | [test name] | [finding] | [Yes/No/Ambiguous] | ### Known Anomalies in This Context - [List relevant anomalies and their current status] ### Conclusion - [Efficiency assessment with caveats] - [Joint-hypothesis caveat]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 23,381 | 13,293 | -43% | 1 | 1 | 0% | 3,902 | 3,111 | -20% | 0 | 0 | — |
case-02 | fail→pass | 40,377 | 12,999 | -68% | 1 | 1 | 0% | 6,635 | 3,224 | -51% | 0 | 0 | — |
case-03 | pass→fail | 58,910 | 15,142 | -74% | 1 | 1 | 0% | 8,268 | 3,457 | -58% | 0 | 0 | — |
case-04 | fail→pass | 46,554 | 18,680 | -60% | 1 | 1 | 0% | 6,474 | 3,431 | -47% | 0 | 0 | — |
case-05 | pass→pass | 22,153 | 14,022 | -37% | 1 | 1 | 0% | 3,135 | 2,845 | -9% | 0 | 0 | — |
case-06 | pass→pass | 21,133 | 14,339 | -32% | 1 | 1 | 0% | 2,827 | 3,122 | +10% | 0 | 0 | — |
case-07 | pass→pass | 18,257 | 12,268 | -33% | 1 | 1 | 0% | 2,834 | 2,717 | -4% | 0 | 0 | — |
case-08 | fail→pass | 20,692 | 13,059 | -37% | 1 | 1 | 0% | 3,220 | 2,897 | -10% | 0 | 0 | — |
case-09 | pass→pass | 23,027 | 21,987 | -5% | 1 | 1 | 0% | 3,479 | 3,830 | +10% | 0 | 0 | — |
case-10 | pass→fail | 21,174 | 14,059 | -34% | 1 | 1 | 0% | 3,418 | 2,899 | -15% | 0 | 0 | — |
case-11 | pass→pass | 15,258 | 15,951 | +5% | 1 | 1 | 0% | 2,367 | 3,420 | +44% | 0 | 0 | — |
case-12 | fail→pass | 17,507 | 12,741 | -27% | 1 | 1 | 0% | 2,471 | 2,897 | +17% | 0 | 0 | — |
case-13 | pass→pass | 17,337 | 14,371 | -17% | 1 | 1 | 0% | 2,567 | 3,031 | +18% | 0 | 0 | — |
case-14 | pass→fail | 21,537 | 19,886 | -8% | 1 | 1 | 0% | 2,924 | 3,558 | +22% | 0 | 0 | — |
case-15 | fail→pass | 18,789 | 17,330 | -8% | 1 | 1 | 0% | 2,971 | 3,506 | +18% | 0 | 0 | — |
case-16 | fail→pass | 19,071 | 14,495 | -24% | 1 | 1 | 0% | 2,883 | 3,125 | +8% | 0 | 0 | — |
case-17 | fail→pass | 23,239 | 14,945 | -36% | 1 | 1 | 0% | 3,518 | 3,240 | -8% | 0 | 0 | — |
case-18 | pass→pass | 15,637 | 15,769 | +1% | 1 | 1 | 0% | 2,326 | 3,138 | +35% | 0 | 0 | — |
case-19 | fail→fail | 21,447 | 13,453 | -37% | 1 | 1 | 0% | 3,288 | 2,929 | -11% | 0 | 0 | — |
case-20 | pass→pass | 20,635 | 31,704 | +54% | 1 | 1 | 0% | 5,003 | 8,244 | +65% | 0 | 0 | — |
case-21 | pass→pass | 8,605 | 10,103 | +17% | 1 | 1 | 0% | 2,184 | 3,407 | +56% | 0 | 0 | — |
case-22 | pass→pass | 9,932 | 10,324 | +4% | 1 | 1 | 0% | 2,270 | 3,321 | +46% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +23 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.